Inhibition of Nuclear Factor-Kappa B Activation Decreases Survival of Mycobacterium tuberculosis in Human Macrophages

Nuclear factor-kappa B (NFκB) is a ubiquitous transcription factor that mediates pro-inflammatory responses required for host control of many microbial pathogens; on the other hand, NFκB has been implicated in the pathogenesis of other inflammatory and infectious diseases. Mice with genetic disruption of the p50 subunit of NFκB are more likely to succumb to Mycobacterium tuberculosis (MTB). However, the role of NFκB in host defense in humans is not fully understood. We sought to examine the role of NFκB activation in the immune response of human macrophages to MTB. Targeted pharmacologic inhibition of NFκB activation using BAY 11-7082 (BAY, an inhibitor of IκBα kinase) or an adenovirus construct with a dominant-negative IκBα significantly decreased the number of viable intracellular mycobacteria recovered from THP-1 macrophages four and eight days after infection. The results with BAY were confirmed in primary human monocyte-derived macrophages and alveolar macrophages. NFκB inhibition was associated with increased macrophage apoptosis and autophagy, which are well-established killing mechanisms of intracellular MTB. Inhibition of the executioner protease caspase-3 or of the autophagic pathway significantly abrogated the effects of BAY. We conclude that NFκB inhibition decreases viability of intracellular MTB in human macrophages via induction of apoptosis and autophagy.


Introduction
Tuberculosis (TB) remains problematic throughout much of the world due, in large part, to burgeoning drug resistance and human immunodeficiency virus (HIV) co-infection [1][2][3]. In addition to these critical barriers to TB control, Mycobacterium tuberculosis (MTB) has been highly successful in evading host immune mechanisms and surviving within phagocytes [4][5][6][7]. Gaining a better understanding of host-pathogen interactions will enable the development of novel therapeutics to treat drug resistant cases and shorten current treatment time.
We examined how inhibition of NFkB activation impacts the viability of intracellular MTB in human macrophages including differentiated THP-1 monocytes, primary monocyte-derived macrophages (MDM), and primary alveolar macrophages (AM). We found that inhibiting NFkB activation reduced the viability of intracellular MTB in all three types of human macrophages. We also examined the impact of inhibiting NFkB activation on apoptosis and autophagy [26][27][28], two cellular processes known to facilitate mycobacterial killing [4,6,[29][30][31][32][33]. We discovered that inhibition of NFkB activation reduces intracellular survival of MTB by enhancing both host-protective apoptosis and autophagy of the infected macrophages.

Differentiation of THP-1 cells
THP-1 cells were cultured in RPMI 1640 medium supplemented with 10% FBS and 2 mM glutamine and were maintained between 2 and 10610 5 cells/mL. Prior to experimentation, THP-1 cells were stimulated with 15 ng/mL PMA for 2 to 3 days to allow differentiation into macrophages. The PMA-containing medium was removed and replaced with fresh medium just prior to exposing the cells to the experimental conditions.

Isolation of human monocyte-derived macrophages
Nine healthy, non-smoking volunteers, 21 to 65 years of age, were recruited for blood donation after National Jewish Health Institutional Review Board (NJH-IRB) approval and written informed consent was obtained from each enrolled subject. Human monocytes were isolated from 50 mL of heparinized blood and process for differentiation into macrophages as we previously described [34]. In brief, the samples were centrifuged at 4006g at room temperature for 25 min. The white buffy coat layer was removed, washed, counted, and resuspended in RPMI medium containing 10% FBS to a concentration of 46 10 6 cells/mL. One-half mL of the cell suspension (2610 6 cells/ 0.5 mL) was added to each well of a 24-well polystyrene plate, estimated to yield about 2610 5 MDM assuming 10% of peripheral white blood cells are monocytes. The cells were incubated at 37uC in a humidified 5% CO 2 incubator for 10-14 days, and the media were replaced on days 2, 5, 7, 9 and 12, resulting in the selection of MDM.

Isolation of alveolar macrophages
Nine healthy, non-smoking volunteers, 21 to 65 years of age, were recruited for bronchoalveolar lavage to obtain AM after NJH-IRB approval and written informed consent was obtained from each enrolled subject. All bronchoscopies were performed by EDC. The bronchoscope was wedged in a segment of the right middle lobe and four-60 mL aliquots of sterile normal saline were instilled and sequentially aspirated back. The volume of lavage recovered was typically 60 to 70% of the amount instilled. The bronchoalveolar lavage fluid was centrifuged at 2006g for 10 min at 4uC. Cell pellets were washed with PBS and resuspended in 10 mL RPMI medium containing 10% FBS and 100 U/mL penicillin G. Cells were counted using a hemocytometer and the volume of medium was adjusted to give a concentration of 1.0610 6 cells/mL. One-quarter mL (2.5610 5 cells) of this suspension plus 250 mL of RPMI medium was added to each well of a 24-well plate and incubated at 37uC in a humidified 5% CO 2 incubator. After 24 hours of incubation, the medium was replaced with fresh antibiotic-free RPMI medium containing 10% FBS and incubated overnight. Prior to infection with MTB, the medium was replaced a second time with antibiotic-free RPMI medium to remove any trace of penicillin G.

Infection of macrophages with MTB
Macrophages were infected with MTB H37Rv at a multiplicity of infection (MOI) of 10 bacilli to 1 macrophage as previously described [35]. In conditions where NFkB was inhibited, the cells were pre-incubated with 5 mM BAY or 0.1% (v/v) DMSO for 1 hr prior to infection.

Culture of cell-associated MTB
For cells infected for 1 hr (Day 0), the supernatants were collected after 1 hr of infection and the cells washed twice with a 1:1 solution of RPMI:16 PBS. Adherent cells were lysed with 250 mL of a 0.25% SDS solution per well, followed by addition of 250 mL of 7H9 plating broth. Serial dilutions of cell lysates were prepared and then 5 mL of each dilution was plated on Middlebrook 7H10 agar. For Day 4 and 8 infections, the cells were washed twice after the initial 1 hr of infection and replaced with fresh RPMI medium containing 10% FBS65 mM BAY; the medium was not changed until 4 or 8 days after infection. After 4 and 8 days of infection, the supernatants were centrifuged to recover any non-adherent macrophages; these macrophages were lysed together with the adherent macrophages and MTB cultured as described above. We considered it important to culture both adherent and non-adherent macrophages to accurately recover all cell-associated MTB. The cell-free supernatants were collected for cytokine analysis.
Construction and detection of MTB H37Rv-GFP MTB H37Rv was transformed with the plasmid pBCM. In brief, the GFP fragment recovered from digestion of plasmid pFPCA1 [36] with BamHI/EcoRI was cloned into pMV261 to make pBCM. pMV261 (Addgene) is a replicative vector that contains a kanamycin resistance gene, an E.coli origin of replication (oriE), a mycobacterial plasmid DNA origin of replication (oriM), an expression cassette containing a mycobacterial promoter (Hsp60), a multiple cloning site, and a transcriptional terminator. After transformation of MTB H37Rv with pBCM by electroporation, positive clones were selected on Middiebrook 7H10 agar which contains 50 mg/mL of kanamycin and GFP-labeled MTB detected by fluorescent microscopy.
THP-1 cells (2610 5 cells/200 mL/well) were plated in black, clear-bottom 96-well microtiter plates in order to minimize background fluorescence (Black Viewplates, Packard Instrument Company, Meriden, Conn). Following overnight differentiation with PMA, the cells were incubated with 5 mM BAY for 1 hr and then infected with MTB H37Rv-GFP at a MOI of 10 for 1 hr. For the 1 hr time point, the medium was removed, 4% paraformaldehyde was added for 1 hr to fix the cells, washed three times with PBS, and fluorescence was measured using a Cytofluor II microplate fluorometer in the bottom-reading mode with excitation and emission wavelengths of 485 nm and 508 nm, respectively. For the 4 and 8 day time points, the medium was changed to fresh medium containing 5 mM BAY, and the MTB-infected cells were incubated at 37uC in 5% CO 2 for an additional 4 and 8 days before the same preparations were made to measure fluorescence. The mean fluorescent value of three wells containing uninfected THP-1 cells was subtracted from the measured fluorescence of all test wells.

Electrophoretic mobility shift assay
NFkB activation was measured by electrophoretic mobility shift assay (EMSA). The human macrophages were either untreated, or pre-treated with 5 mM BAY or 0.1% DMSO vehicle for 1 hr, followed by infection with MTB H37Rv at a MOI of 10 for 30 min, 1 hr, 3 hrs, or 6 hrs. The methods used for the nuclear protein extraction, end-labeling of a double stranded oligonucleotide (Promega, Madison, WI) containing a consensus sequence for NFkB with 32 P-c-ATP, and binding reaction of the nuclear proteins with the labeled oligonucleotide are as previously published [37,38]. One additional step was that the extracted nuclear protein was filtered through a 0.2 mm syringe filter to remove any MTB.

Cytokine measurements
The concentrations of TNFa and IL-8 in the culture supernatants were determined by ELISA according to manufacturer's instructions. The level of IFNc was determined by electrochemiluminescence as previously described [35,39].

Apoptosis assay
Apoptosis was determined by terminal deoxynucleotidyl transferase dUTP nick end labeling (TUNEL). Cell suspensions (0.5 mL) were grown on 4-chamber well slides [34], pre-incubated with BAY and then infected with MTB H37Rv at a MOI of 10. The medium was removed at 4 and 8 days following infection and the cells were fixed in 4% paraformaldehyde solution (pH 7.4). The cells were then stained with an Apoptosis In Situ Detection Kit (Roche Diagnostic, Indiannapolis, IN) according to the manufacturer's instructions and as previously reported [35]. Enumeration of apoptotic cells was performed by light microscopy, counting ,5000 cells per condition per experiment.

Western blotting
Expression of the autophagosome-associated LC3-II and its precursor LC3-I as well as cytochrome c were detected by western blotting using methods previously described [35]. In brief, after SDS-PAGE of nuclear-free whole cell lysates (20 mg protein per condition), the proteins were transferred onto a nitrocellulose membrane (ISC BioExpress, Kaysville, UT), blocked in 10% (w/v) non-fat dry milk at 4uC overnight with shaking, and immunoblotted with 1:1000 solution of anti-LC3B antibody in 5% BSA (w/v) or anti-cytochrome c antibody in 5% (w/v) non-fat dry milk for 1 hr at room temperature shaking. After incubation with horseradish peroxidase-conjugated secondary anti-rabbit antibody (1:2000) in 5% (w/v) non-fat dry milk for 1 hr at room temperature, bound antibodies were detected by enhanced chemoluminescence.

Direct fluorescence to examine autophagosome formation
Autophagosomes can be directly visualized as cytoplasmic granules (punctae) that stain positive for LC3-II, a lipoprotein specific for autophagosomal membranes [40]. To quantify LC3-IIpositive punctae in MTB-infected THP-1 cells with NFkB inhibition, recombinant GFP-LC3 lentivirus was generated in the HEK 293 cell line (Cell Biolabs, Inc, San Diego, CA). Lentiviruses were prepared by co-transfecting packaging plasmids pVSV-g (Dr. Jerome Schaack, University of Colorado School of Medicine) and pCMV-HIVdeltaR8.2 (Addgene, plasmid #12263) using an established protocol [41]. The GFP-LC3 expression plasmid, pCDH-GFP-LC3, was generated by subcloning the GFP-LC3 gene from the pEGFP-C3 vector backbone (Addgene) into the pCDH-puro lentivirus expression vector. GFP-LC3 lentivirus was transduced into THP-1 cells at a multiplicity of 10 lentiviruses to one THP-1 cell in the presence of 8 mg/mL Polybrene; GFP lentivirus served as a positive control and empty vector lentivirus as a negative control. To quantify autophagosome formation by fluorescence, GFP-LC3 lentivirus-transduced cells were plated on chamber slides at 10 5 cells per well, and allowed to differentiate overnight with 15 ng/mL PMA at 37uC in 5% CO 2 . The next day, the PMA-containing medium was replaced with fresh THP-1 medium (RPMI+10% FBS+1% glutamine). BAY was added at a concentration of 5 mM per well and was incubated for 1 hr prior to infection with MTB. Cells were infected with MTB H37Rv at a MOI of 10 for 24 hrs and fixed in a 4% paraformaldehyde solution for 1 hr at room temperature. After rinsing with PBS, the cells were then stained with ProLong Gold Antifade reagent with DAPI (Invitrogen) and placed in the dark at 4uC until use. For quantifying the number of autophagosomes per cell, the number of GFP-LC3 punctae were counted from at least 100 cells/well in duplicate wells for each condition. All the cellular images were obtained using an inverted confocal fluorescence microscope (Nikon, Carl Zeiss Axiovert 200 M) and the data analysis using correlation analysis tool of Slidebook program (Slidebook 4.2).

Statistical analysis
Replicate experiments were independent, and where appropriate, summary results are presented as means 6 SEM. Differences were considered significant for p,0.05, and all reported p-values used a two-sided test. For most experiments, group means were compared by ANOVA using Fisher's least significance difference procedure. The AM and MDM cytokine data were analyzed as follows to account for repeated measures over treatment and time within subjects. Each outcome variable was fit with a linear mixed model, using predictor variables of MTB (Y, N), BAY (Y, N), Time (1, 24, 96, 192 hours), and all interactions of these, up to the 3-way interaction. Repeated measures on subjects over treatment (4 combinations of MTB-Y/N and BAY-Y/N) and time were accounted for using a direct (Kronecker) product structure for all outcome variables except for those noted below. The ''unstructured'' covariance structure was used for repeated measures involving time, while the compound symmetry structure was used for those involving treatment; the direct product structure allows these to be modeled simultaneously in one combined covariance structure. TNFa and IFNc produced by MDM used a simpler compound symmetric covariance structure since for these outcome variables the models did not converge using the direct product covariance structure.
Outcome variables were log transformed for analysis when necessary (AM-TNFa, THP-1-TNFa, and AM-IFNc), but results for all outcome variables are presented on the original scale; those analyzed on the log scale were inverted back for presentation.
Specific pairwise comparisons of interest were between times for the MTB only treatment, and within times between the MTB only and MTB + BAY treatments. Such comparisons were only considered when at least one predictor variable (main effect or interaction) in the related model was significant.

MTB-induced NFkB activation is inhibited by BAY
MTB induced NFkB activation in THP-1 cells, particularly after 3 and 6 hrs of infection ( Figure 1A). In the presence of 5 mM BAY, NFkB activation was strongly inhibited ( Figure 1A). Coincubation of the nuclear extract with anti-p50-NFkB antibody further retarded the migration of NFkB complexes, previously identified to be p50/p50 and p50/p65 [38]. There was no further reduction in NFkB activation following administration of 7.5 mM or 10 mM BAY (data not shown).
We pre-incubated primary human MDM and AM with 5 mM BAY or 0.1% (v/v) DMSO vehicle for 1 hr, followed by infection with MTB H37Rv for an additional 3 hrs. Similar to THP-1 cells, MTB infection of primary MDM and AM induced NFkB activation, and this activation was inhibited by BAY ( Figure 1B/C).

Inhibition of NFkB significantly reduces the viability of intracellular MTB
Differentiated THP-1 cells, MDM, and AM were used to determine the impact of NFkB inhibition on intracellular recovery of viable MTB. The mean number of intracellular MTB isolated 1 hr following infection (Day 0) was similar between the control cells and cells pre-treated with BAY, indicating that inhibition of NFkB did not impact the phagocytosis of MTB by the macrophages (Figure 2A,B,C). However, inhibition of NFkB activation significantly reduced the number of intracellular MTB recovered 4 days after infection from THP-1 cells by 64%, from MDM by 67%, and from AM by 63% (Figure 2A,B,C). By 8 days after infection, BAY significantly reduced the number of viable intracellular MTB by 66% in THP-1 cells, 63% in MDM, and 71% in AM (Figure 2A,B,C). There was no significant difference in the number of CFU recovered with higher concentrations of BAY (7.5 and 10 mM) compared to 5 mM BAY (data not shown). In the absence of macrophages, 5 mM or 10 mM BAY had no effect on viability of MTB H37Rv compared to MTB cultured with 0.1% DMSO vehicle ( Figure S1).
Trypan blue staining was performed on control THP-1 cells and cells incubated for 4 days with 5 mM BAY or BAY + MTB. Approximately 11% of the cells infected with MTB alone were non-viable whereas ,20% of the cells were non-viable after incubation with both MTB and BAY ( Figure S2).
While we attempted to recover any non-adherent cells at the Day 4 and Day 8 time points since it would be important to include any live MTB in these cells, some cells may have been lost in the recovery process. To validate the CFU data, we infected THP-1 cells with MTB H37Rv-GFP in the presence or absence of BAY at the aforementioned times of infection. After the indicated times, the cells were washed, fixed, and fluorescence intensity measured. As shown in Figure 2D, the fluorescent intensity of the BAY-treated THP-1 cells was significantly less than untreated MTB-infected cells at 4 and 8 days after infection. Moreover, the relative slope of the fluorescent curves roughly matched those of the CFU growth curves.
While BAY does not inhibit other signaling pathways such as the mitogen-activated protein kinases [42,43], we validated our findings by transducing THP-1 cells with an EIA-deficient adenovirus (AdV) construct containing a mutant IkBa (AdV-S32/ 36A-IkBa) [44,45]. Due to lack of EIA, this attenuated AdV-S32/ 36A-IkBa vector is unable to replicate but because the mutated IkBa cannot be phosphorylated, this dominant-negative IkBa remains a potent inhibitor of NFkB activation. We have previously used this construct to effectively inhibit MTB lipoarabinomannan activation of NFkB [38]. THP-1 cells were transduced with AdV-GFP (control vector) or AdV-S32/36A-IkBa for 6 hrs, infected with MTB at a MOI of 10 for 4 days, and cell-associated MTB were quantified. As shown in Figure 3A, AdV-GFP-infected cells had similar numbers of MTB compared to MTB infection of wildtype THP-1 cells; in contrast, the AdV-S32/36A-IkBainfected cells had significantly reduced number of cell-associated MTB, consistent with the results of NFkB inhibition by BAY. To confirm AdV-S32/36A-IkBa is biologically active in THP-1 cells, we determined whether MTB-induction of IL-8 could be inhibited since induction of IL-8 is critically dependent on NFkB [46,47]. THP-1 cells infected with MTB for 24 hrs induced IL-8 but coincubation with BAY significantly inhibited IL-8 expression ( Figure 3B). THP-1 cells transduced with Adv-GFP and infected with MTB also had robust induction of IL-8; in contrast, THP-1 cells transduced with AdV-S32/36A-IkBa and infected with MTB had significantly lower amounts of MTB-induced IL-8 compared to AdV-GFP transduced cells ( Figure 3B).

Inhibition of NFkB increases apoptosis of macrophages
Apoptosis contributes to the intracellular killing of mycobacteria [30,[48][49][50][51]; thus, the effect of NFkB inhibition on apoptosis of infected macrophages was evaluated. Following 4 days of incubation, there was no difference in extent of TUNEL staining between uninfected THP-1 cells and THP-1 cells infected with MTB ( Figure 4A). However, there was a significant increase in apoptosis at 4 days (relative increase in apoptosis of 79%) in BAYtreated, MTB-infected THP-1 cells ( Figure 4A). After 8 days of infection, there was a modest but significant increase in apoptosis with MTB infection alone compared to uninfected cells (relative increase in apoptosis of 34%); in the presence of BAY, there was a further significant increase in apoptosis of MTB-infected THP-1 cells ( Figure 4A).
BAY alone was also found to modestly increase apoptosis of THP-1 cells in the absence of MTB at 24 hrs; i.e., the level of TUNEL positivity in THP-1 cells alone was 2.45%; with the addition of 5 mM BAY, the percent of TUNEL positivity increased to 5.37%. These results suggest that basally, low levels of NFkB activation may prevent spontaneous apoptosis in differentiated THP-1 cells.
We next quantified TUNEL positivity in the MTB-infected MDM and AM in the presence or absence of BAY. MTB-infected MDM and AM showed increased apoptosis after 4 days of infection in the presence of BAY, with relative increases of 170% and 60%, respectively ( Figure 4B,C).
To corroborate that apoptosis is induced by MTB and BAY, we analyzed the level of cytochrome c, an essential component of the intrinsic apoptotic pathway, in THP-1 cells infected with MTB with and without BAY after 48 hrs. MTB alone had little or no induction of cytochrome c as assessed by western blotting ( Figure 4D). BAY 11-7082 treatment alone had a modest induction of cytochrome c whereas culture of THP-1 cells with both MTB and BAY showed a significant induction of cytochrome c ( Figure 4D).
While the absolute change in apoptosis was small, there was a large decrease in bacterial recovery from macrophages. This phenomenon can occur when killing mechanisms other than apoptosis are involved but also when macrophages are nonuniformly infected with MTB. To assess the latter, THP-1 cells were infected with MTB H37Rv-GFP for 1 hr and visualized by fluorescent microscopy. Shown in lower magnification is a representative differential interference contrast (DIC) image ( Figure 4E, panel 1). In a representative fluorescent image, one macrophage is heavily infected (arrows in panels 1 and 2 of Figure 4E) while others were much less burdened or not infected ( Figure 4E, panel 2). The cells were also stained with DAPI and the merged images are shown in Figure 4E, panel 3 and further magnified in Figure 4E, panel 4. Similar findings were also seen with the infected MDM and AM with Auramine acid-fast staining (data not shown).

Caspase-3 inhibition abrogates BAY-induced apoptosis
To determine whether BAY-induced apoptosis impacted viability of intracellular MTB, we analyzed the effect of inhibiting apoptosis using a specific caspase-3 inhibitor (z-DEVD-fmk). Caspase-3 is an executioner caspase for both the intrinsic and extrinsic pathways of classical apoptosis. THP-1 cells infected with MTB or cultured with BAY for 48 hrs showed a modest induction of activated caspase-3 whereas cells cultured with both MTB and BAY showed robust caspase-3 activation ( Figure 5A). Co-culture with 10 mM of the caspase-3 inhibitor (z-DEVD-fmk) showed a significant abrogation of MTB+BAY activation of caspase-3 ( Figure 5A). THP-1 cells were infected with MTB alone, MTB+5 mM BAY, or MTB+BAY and 10 mM z-DEVD-fmk. At 1 hr, 2 days, and 4 days after infection, cells were lysed and lysates cultured for MTB. As seen previously, compared to MTB alone, addition of BAY significantly reduced the number of MTB in THP-1 cells ( Figure 5B). With caspase-3 inhibition, BAYmediated decrease in CFU was significantly abrogated ( Figure 5B).
To confirm that caspase-3 inhibition reduces the induction of apoptosis by MTB and BAY, infected THP-1 cells were incubated with BAY in the presence or absence of the specific caspase-3 inhibitor, z-DEVD-fmk. Increased apoptosis was observed in BAY+MTB-infected cells after 48 hrs; in the presence of z-DEVDfmk, MTB plus BAY induced apoptosis was significantly suppressed ( Figure 5C). There was no significant difference between the effects of 10 mM and 20 mM of z-DEVD-fmk in regard to both intracellular number of MTB and apoptosis (data not shown).

Inhibition of NFkB induces autophagy in MTB-infected THP-1 cells
Autophagy is a cellular process by which cytoplasmic organelles and molecules are catabolized, and is increasingly recognized as an effector mechanism by which host cells kill intracellular MTB and other intracellular pathogens [4,6,52]. With initiation of autophagy, an LC3-II-mediated double membrane forms around the cytoplasmic or intraphagosomal organisms, forming autophagosomes that more readily fuses with lysosomes, resulting in the killing of the organisms by lysosomal machinery [40].
Immediately following synthesis of LC3, its C-terminal fragment is cleaved to yield a cytosolic form of LC3 termed LC3-I. During induction of autophagy, LC3-I is conjugated with phosphatidylethanolamine (PE) to form LC3-II, which binds to both the outer and inner autophagosomal membranes [53]. Thus, LC3-II is a highly specific marker of autophagosome formation [6]. Although LC3-II has a higher molecular weight than LC3-I, it runs lower on SDS-PAGE due to the physical properties of the lipid attachment. Thus, we determined the effects of NFkB inhibition and MTB infection on the levels of the autophagic marker LC3-II.
THP-1 cells were exposed to the following conditions for 24 hrs: RPMI medium+10% FBS+1% glutamine+0.1% DMSO (unstimulated control), RPMI without FBS (starvation control), 5 mM BAY, MTB, or both MTB and BAY, followed by western blotting for LC3-I and LC3-II. This time point were found to be optimal to observe an increase in LC3-II by autophagy induction. Immunoblot showed increased LC3-II in lysates of starved cells compared to unstimulated control cells (DMSO) ( Figure 6A). BAY treatment or MTB infection also modestly induced more LC3-II formation and when the cells were incubated with both MTB and BAY, there was a further increase in LC3-II ( Figure 6A).
To validate the immunoblot data, we quantified the average number of LC3-II-positive autophagosomes per cell using THP-1 cells transduced with GFP-LC3 lentivirus. Transduced THP-1 cells were either left uninfected or infected with MTB for 24 hrs. We found that MTB infection alone significantly increased the number of GFP-LC3 punctae per cell compared to uninfected cells by 4 to 5-fold ( Figure 6B). Following NFkB inhibition of MTBinfected THP-1 cells, the average number of GFP-LC3 punctae per cell further increased by an additional ,2.5-fold ( Figure 6B).
To corroborate whether autophagy plays a role in reducing the number of viable intracellular MTB recovered, we inhibited autophagy in MTB-infected THP-1 cells that were treated with BAY using the established autophagy inhibitor 3-methyladenine (3-MA), an inhibitor of class III PI3 kinase, an upstream signaling molecule of the autophagic signaling cascade [54]. As shown in Figure 6C, LC3-II was increased in MTB-infected THP-1 cells treated with BAY compared to THP-1 cells infected with MTB alone. With addition of 6 mM 3-MA in the culture medium, there was significant reduction in LC3-II levels. THP-1 cells were then infected with MTB alone or MTB+5 mM BAY with and without addition of 6 mM 3-MA. The cells were then cultured for 4 days and cell-associated MTB was quantified. As shown in Figure 6D, BAY reduced the number of MTB recovered, which was significantly abrogated by the addition of 3-MA. The effects of NFkB inhibition on intracellular MTB are independent of TNFa and IFNc expression TNFa can induce apoptosis [31,55,56] and IFNc can induce autophagy in MTB-infected macrophages through immunityrelated p47 guanosine triphosphatases (IRG) [6,33]. Additionally, NFkB is an important transcription factor in the induction of both IFNc and TNFa [57]. We examined the production of TNFa and IFNc in MTB-infected human macrophages with or without NFkB inhibition. THP-1 cells had no detectable IFNc and had minimal induction of TNFa by MTB at the time points examined (data not shown). In MDM, TNFa and IFNc were minimally induced by MTB infection (Figure 7). In AM, TNFa and IFNc were induced by MTB infection and BAY significantly inhibited their expression (Figure 7). The inhibition of TNFa and IFNc expression in AM by BAY occurred despite concomitant decreases in MTB recovered. These results indicate that inhibition of bacterial growth following inhibition of NFkB activation is likely independent of TNFa and IFNc production.

Discussion
In human macrophages, inhibition of NFkB activation reduced the viability of intracellular MTB through increased induction of apoptosis and autophagy. While we showed that inhibition of NFkB enhanced macrophage response to MTB, it is important to note that NFkB is a ubiquitous transcription factor involved in many cellular processes associated with inflammation and infections. For example, suboptimal binding of NFkB to a cisregulatory site on the 59-flanking region of the IFNc gene promoter leads to decreased production of IFNc [58], a cytokine that is critically important in host-defense against mycobacterial infections. While others have shown that NFkB activation in mouse macrophages resulted in increased killing of mycobacteria, those studies primarily used non-pathogenic Mycobacterium smegmatis [59,60]. These studies, combined with our findings, suggest that NFkB plays roles in both host-defense and host-susceptibility, depending on the microbial pathogen and the host species.
Since mice lacking the p50 subunit of NFkB suffered worsened MTB infection as compared to wildtype mice [23], we predicted that pharmacologic inhibition of NFkB activation in human macrophages would result in increased recovery of viable intracellular MTB. In contrast, we found that inhibition of NFkB activation resulted in significantly fewer viable intracellular MTB despite similar number of bacilli phagocytosed. One possible reason for the discrepancy between the prior mouse study and our current work with human macrophages is that production of nitric oxide, which is critically dependent on NFkB activation, plays a central role in killing intracellular MTB in mice [61]. Indeed, MTB-infected p50 -/-mice had significantly less inducible nitric oxide synthase expression in their lung tissues than infected wildtype mice [23]. In contrast, the role of nitric oxide in killing MTB in humans and human macrophages is considered far less important [62]. Additionally, there are inherent differences in the in vivo mouse study and our in vitro/ex vivo human macrophage studies. For example, in the p50-knockout mice, NFkB complexes containing the p50 subunit are silenced in all cell types including macrophages; in contrast, we examined macrophages in isolation. On the other hand, because we used a pharmacologic inhibitor of IKK as well as a dominant-negative IkBa, our experimental approach is expected to inhibit activation of not only p50containing NFkB complexes but all NFkB complexes, which are comprised of dimers of various subunits (Rel A, c-Rel (p65), RelB, p50, and p52). Since there are several different types of NFkB complexes, an additional possibility is that there are differences between mouse and human cells in the ability of various NFkB complexes to affect immune response. This notion is consistent with the work of Gutierrez and co-workers who showed that in murine macrophages, activation of NFkB by some host lipids can induce killing of MTB whereas other lipids can induce killing of mycobacteria independent of NFkB activation [60]. Finally, the mouse study employed a different strain of MTB (Kurono strain) than the current study (H37Rv). Since different strains of MTB can induce varying degrees of apoptosis in host cells [51], differences in the MTB strain used may impact the importance of apoptosis in killing intracellular MTB with NFkB inhibition.
After discovering that inhibition of NFkB activation decreased intracellular MTB recovery, we next sought out potential mechanism(s). We found in all three types of human macrophages that decreased number of MTB recovered following NFkB inhibition correlated with increased apoptosis, an important effector killing mechanism of intracellular MTB [30,48,50,51,53]. Moreover, since there was non-uniformity of MTB infection of macrophages, apoptosis of relatively few but heavily infected cells could account for the significant reduction in the number of viable intracellular MTB recovered. While inhibition of classical apoptosis with inhibition of caspase-3 significantly abrogated the effects of BAY in terms of CFU and apoptosis, the reversal was not complete, suggesting that the effects of BAY may also involve caspase-3-independent programmed cell death mechanisms [63]. Based on our finding that the percent of non-viable cells as determined by Trypan blue staining was greater than the percentage of cells that were undergoing apoptosis, we believe that some of the infected cells are also undergoing necrosis as well. This is supported by studies showing that with MOI of 5 and 10, MTB H37Rv caused both apoptosis and necrosis of human macrophages [64]. This may explain the finding that even though inhibition of NFkB decreased the number of cellassociated MTB, the growth curve was still increasing over time; i.e., we posit that in contrast to cells undergoing apoptosis, MTB remain viable in necrotic cells and upon release, have the potential to be phagocytosed by other macrophages and proliferate. Thus, we believe the overall fate of MTB cultured in macrophages is a dynamic process with some bacteria being killed while others proliferate, with the net biological effect being decreased number of intracellular MTB with inhibition of NFkB.
While TNFa has been shown by others to induce apoptosis in the context of MTB infection, there was little induction of TNFa in either THP-1 cells or MDM infected with MTB. For the THP-1 cells, we believe the minimal TNFa production may be due to the development of ''tolerance'' of these cells to subsequent MTBinduced cytokine production since we incubated the THP-1 cells with PMA for 2 to 3 days before changing the medium and infecting with MTB. We have since found that differentiating these cells over a shorter, overnight PMA incubation preserves the ability of THP-1 cells to produce cytokines in response to MTB without impact on intracellular survival of MTB [35].
Dhiman and co-workers previously showed that THP-1 cells, in which NFkB was suppressed via transfection with a mutated, dominant-negative form of IkBa, were less burdened by intracellular MTB [65]. Similar to our findings, the decrease in bacterial growth due to suppression of NFkB coincided with increased apoptosis of the transfected THP-1 cells [65]. We extended these findings by utilizing two types of primary human macrophages. Additionally, we showed that inhibition of classical apoptosis abrogated the reduction in intracellular CFU induced by NFkB inhibition. While Loeuillet et al showed that TLR2-mediated activation of NFkB prevented MTB infected THP-1 cells from undergoing Fas ligand mediated apoptosis [66], the anti-apoptotic effect of NFkB may involve induction of several anti-apoptotic gene products that ultimately inhibit caspase-3 activation [28].
NFkB may also interfere with apoptosis via direct protein-protein interaction such as direct coupling of NFkB subunits and c-IAP2, providing a signal amplification loop that promotes cell survival independent of de novo protein synthesis [28]. Activation of NFkB in the context of infections with Bartonella, Ehrlichia, or Rickettsia has also been shown to inhibit apoptosis of host cells by preventing the release of cytochrome c [67]. The increased vulnerability of AIDS patients to TB is most likely related to impaired effector T cell function [68,69]. However, increased NFkB activation seen in HIV positive individuals could also impair the ability of their MTB-infected macrophages to undergo apoptosis [56,70], providing another mechanism for their predisposition to TB.
NFkB inhibition also increased the formation of autophagosomes. Since autophagy has been shown to be an efficient killing mechanism of intracellular MTB [4,6,52], induction of apoptosis and autophagy are two mechanisms by which NFkB inhibition reduced the number of intracellular bacilli (Figure 8). Due to experimental limitations, we could not determine whether apoptotic cells were the exact same cells undergoing autophagy. However, it is plausible that autophagy may be activated in dying cells, perhaps as an attempt by dying cells to survive [71]. Since we detected increased autophagy in MTB-infected cells in which NFkB activation was inhibited at 24 hrs, whereas induction of apoptosis was seen later, it suggests that autophagy is not likely to be a secondary response to apoptosis in our experimental model. However, we cannot exclude the possibility that autophagy may be induced in response to the earliest stages of apoptosis. Consequently, there is a possibility that both processes may occur in the same cell over an extended period of time, but experimentally this would be very difficult to prove. Furthermore, the interplay between these NFkB-mediated mechanisms of MTB killing is likely complicated by the fact that NFkB can regulate its own activation by opposing mechanisms; i.e., NFkB activation induces the production of its inhibitory molecule IkBa (Figure 8, blue arrow) and yet NFkB inhibition of autophagy can increase IKK activity since this kinase is normally degraded by the autophagic process [72] (Figure 8, red line).
Cytokines and microbial products can activate other transcription factors such as AP-1 and ATF-2, which likely play important roles in the host immune response to MTB [42,73]. A class of transcription factors known as nuclear receptors have recently been implicated in host-mycobacterial interactions [74,75]. Mahajan and co-workers [75] showed that during MTB infection of macrophages, MTB-derived lipids and macrophage-derived lipids can combine with lipid-sensing nuclear receptorsperoxisome proliferator-activated receptor gamma (PPARc) and testicular receptor 4 (TR4) -to induce expression of genes that ultimately enhances intracellular survival of MTB. In addition, activation of either of these receptors induced alternate activation of macrophages to the M2 phenotype [75], which would be expected to impair effective host immune response against MTB. These nuclear receptors are immune evasive factors since mycobacteria can induce the expression of PPARc [74]. In contrast, activation of another type of nuclear receptor (LXRa) decreases lipidogenesis and enhances host immunity against MTB [75].
In summary, we found that inhibiting NFkB activation in macrophages resulted in increased apoptosis and autophagy, and decreased recovery of viable intracellular MTB. There are hundreds of natural and synthetic compounds known to inhibit NFkB activation, including various antioxidants, proteasome and protease inhibitors, and IKK inhibitors (http://people.bu.edu/ gilmore/nf-kb/inhibitors/index.html). It is clear that the role of NFkB following MTB infection is complicated. Future studies could consider utilizing mixed cell cultures to determine the effects of NFkB inhibition on the collaboration between macrophages and T cells. Furthermore, it is plausible that NFkB activation may be important in the early phase of infection but continued activation may be deleterious to the host. Thus, future in vivo studies could inhibit NFkB at different time points of infection to begin to delineate ''host-protective'' vs. ''host-susceptible'' NFkBmediated pathways.